Number of write cycles to STM32F030 flash memory
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 08:32 AM
I am using the STM32F030, what is the minimum number of write cycles i can make to flash memory for my configuration data?
Also what would it be for the STM32L073RZ?
Thanks
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 08:44 AM
This is something you should be able to find in the Data Sheet
For the STM32L073RZ, Flash endurance min 10K cycles, EEPROM min 100K cycles, under nominal conditions
If you journal the write of your configuration data structure over the sector/block rather than erasing/writing each time, you can extend the life. For something on a battery all the time, consider if the use of NVRAM/BKPRAM would be a wiser choice if the frequency of change is high.
Up vote any posts that you find helpful, it shows what's working..
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 08:50 AM
Look into the datasheet:
Page 62 : Table 40. Flash memory endurance and data retention
An artificial way to increase this number by a factor of n is to use n times the size of the configuration data as the number of cycles should be related to erase cycles, or use an external I2C EEPROM to get very high cycling number.
Hope this helps.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 08:56 AM
Hello
Barningham.Andrew
‌,The information aboutflash memory can be found in thedatasheet of your product.
So, look at the following:
- STM32F030 datasheet in the
Memory characteristics section.
- STM32L073RZ
:Table Flash memory and data EEPROM endurance and retention , inthe
Memory characteristics section.
Best Regards
Imen
Thanks
Imen
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 08:57 AM
The
http://www.st.com/resource/en/datasheet/stm32f030rc.pdf
suggests 1000 cycles - which seems quite a few. Note that this value is guaranteed over the whole temperature range, and 'based on characterization, not production testing'.Actual number of possible cycles usually increases exponentially with lower temperatures (min. 1kcycles is at +85°C).
10 kCycles for the STM32F072 and STM32F073.
Finding/reading datasheets is an essential skill for embedded engineers ...
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 09:49 AM
Thanks... i did see this in the Datasheet but thought it was low so wanted to double check!!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 09:59 AM
Thanks, yes that does helps.
Page 62.... 1K.. this is very low!
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 10:03 AM
That's the price for the price... The STM32F030 is a stripped-down-to-the-bone-for-minimum-cost member of the 'F0x subfamily. That stripdown includes minimized production testing. It had been
https://community.st.com/0D50X00009XkigqSAB
as 'https://web-beta.archive.org/web/20130909234338/http://www.st.com/web/en/press/en/p3444
'.JW
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 10:21 AM
1K is rather grim, I wonder how the array/geometry differs from the 10K rated designs to get down to that....
My guess would be the other designs have extra bits in the wide flash lines committed to ECC, so a single bit doesn't result in a hard failure.
Up vote any posts that you find helpful, it shows what's working..
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Email to a Friend
- Report Inappropriate Content
‎2016-12-21 10:27 AM
Not sure the 1K vs 10K impacts tester time, unless they don't do at least one erase/write/erase cycle, because they say the number comes from characterization of the silicon. ie they destructively tested several batches of devices from corners of the process window, and the failures lead them to be comfortable with specifying only 1K cycles.
Up vote any posts that you find helpful, it shows what's working..